In [1]:
save_augmented_data = False
Step 0: Load The Data¶
In [2]:
# Load pickled data
import pickle
# TODO: Fill this in based on where you saved the training and testing data
training_file = 'train.p'
validation_file= 'valid.p'
testing_file = 'test.p'
with open(training_file, mode='rb') as f:
train = pickle.load(f)
with open(validation_file, mode='rb') as f:
valid = pickle.load(f)
with open(testing_file, mode='rb') as f:
test = pickle.load(f)
X_train, y_train = train['features'], train['labels']
X_valid, y_valid = valid['features'], valid['labels']
X_test, y_test = test['features'], test['labels']
use_n_channels = 3
Step 1: Dataset Summary & Exploration¶
The pickled data is a dictionary with 4 key/value pairs:
'features'is a 4D array containing raw pixel data of the traffic sign images, (num examples, width, height, channels).'labels'is a 1D array containing the label/class id of the traffic sign. The filesignnames.csvcontains id -> name mappings for each id.'sizes'is a list containing tuples, (width, height) representing the original width and height the image.'coords'is a list containing tuples, (x1, y1, x2, y2) representing coordinates of a bounding box around the sign in the image. THESE COORDINATES ASSUME THE ORIGINAL IMAGE. THE PICKLED DATA CONTAINS RESIZED VERSIONS (32 by 32) OF THESE IMAGES
Provide a Basic Summary of the Data Set Using Python, Numpy and/or Pandas¶
In [3]:
### Replace each question mark with the appropriate value.
### Use python, pandas or numpy methods rather than hard coding the results
# TODO: Number of training examples
n_train = len(X_train)
# TODO: Number of validation examples
n_validation = len(X_valid)
# TODO: Number of testing examples.
n_test = len(X_test)
# TODO: What's the shape of an traffic sign image?
image_shape = X_train[0].shape
# TODO: How many unique classes/labels there are in the dataset.
n_classes = len(set(y_train))
print("Number of training examples =", n_train)
print("Number of validation examples =", n_validation)
print("Number of testing examples =", n_test)
print("Image data shape =", image_shape)
print("Number of classes =", n_classes)
Include an exploratory visualization of the dataset¶
In [4]:
import csv
with open('signnames.csv', mode='rt') as file:
file.readline() # discard header
sign_names = { int(r[0]) : r[1] for r in csv.reader(file) }
Original Data Preview¶
In [5]:
import matplotlib.pyplot as plt
import numpy as np
import math
%matplotlib inline
classes, indices = np.unique(y_train, return_index=True)
columns = 5
rows = math.ceil(n_classes/columns)
figsize=(16, 16)
fontsize=8.5
fig, axes = plt.subplots(rows, columns, figsize=figsize,
subplot_kw={'xticks': [], 'yticks': []})
fig.subplots_adjust(hspace=0.3, wspace=0.10)
for ax, i in zip(axes.flat, indices):
ax.imshow(X_train[i], interpolation='lanczos', cmap='viridis')
ax.set_title(sign_names[y_train[i]], fontsize=fontsize)
plt.show()
In [6]:
from collections import Counter
fig, ax = plt.subplots(figsize=(20,5))
D = Counter(y_train)
plt.xticks(range(len(D)), D.keys())
plt.bar(range(len(D)), D.values(), color="red", alpha=0.8, label='Training')
D = Counter(y_test)
plt.xticks(range(len(D)), D.keys())
plt.bar(range(len(D)), D.values(), color="blue",label='Testing')
D = Counter(y_valid)
plt.xticks(range(len(D)), D.keys())
plt.bar(range(len(D)), D.values(), color="green",label='Validation')
legend = ax.legend()
Step 2: Design and Test a Model Architecture¶
Design and implement a deep learning model that learns to recognize traffic signs. Train and test your model on the German Traffic Sign Dataset.
Here is an example of a published baseline model on this problem. It's not required to be familiar with the approach used in the paper but, it's good practice to try to read papers like these.
Pre-process the Data Set¶
Minimally, the image data should be normalized so that the data has mean zero and equal variance. For image data, (pixel - 128)/ 128 is a quick way to approximately normalize the data and can be used in this project.
Other pre-processing steps are optional. You can try different techniques to see if it improves performance.
In [7]:
from sklearn.utils import shuffle
import tensorflow as tf
def preprocess_gpu(x_in):
x_t = tf.cast(x_in, dtype=tf.float32)
x_out = []
if use_n_channels == 1:
x_t = tf.image.rgb_to_grayscale(x_t)
x_t = tf.map_fn(lambda img: tf.image.per_image_standardization(img), x_t, dtype=tf.float32)
with tf.Session() as sess:
x_out = x_t.eval()
return x_out
In [8]:
%%time
X_train = preprocess_gpu(X_train)
X_valid = preprocess_gpu(X_valid)
X_test = preprocess_gpu(X_test)
Prepocessed Data Preview¶
In [9]:
fig, axes = plt.subplots(rows, columns, figsize=figsize,
subplot_kw={'xticks': [], 'yticks': []})
fig.subplots_adjust(hspace=0.3, wspace=0.10)
for ax, i in zip(axes.flat, indices):
ax.imshow(X_train[i], interpolation='lanczos', cmap='viridis')
ax.set_title(sign_names[y_train[i]], fontsize=fontsize)
plt.show()
In [10]:
import math
def distort_input_gpu(x_in):
x_t = tf.cast(x_in, dtype=tf.float32)
x_t = tf.map_fn(lambda img: tf.random_crop(img, [32, 32, use_n_channels]), x_t, dtype=tf.float32)
x_t = tf.map_fn(lambda img: tf.image.random_brightness(img, max_delta=63./255.), x_t, dtype=tf.float32)
x_t = tf.map_fn(lambda img: tf.image.random_saturation(img, lower=0.5, upper=1.5), x_t, dtype=tf.float32)
x_t = tf.map_fn(lambda img: tf.image.random_hue(img, max_delta=0.2), x_t, dtype=tf.float32)
x_t = tf.map_fn(lambda img: tf.image.random_contrast(img, lower=0.2, upper=1.8), x_t, dtype=tf.float32)
x_t = tf.map_fn(lambda img: tf.image.per_image_standardization(img), x_t, dtype=tf.float32)
with tf.Session() as sess:
x_out = x_t.eval()
return x_out
In [11]:
def augment_class(x_in, y_in, classid, target):
x_augment = np.empty([0,32, 32, 3])
y_augment = []
running_total = 0
selected_class = np.where(y_in == classid)
x_class_agument = []
while running_total < target:
for i in selected_class[0]:
x_class_agument.append(x_in[i])
y_augment.append(classid)
running_total += 1
if running_total >= target:
break
distorted_x = distort_input_gpu(np.array(x_class_agument))
x_augment= np.append(x_augment, distorted_x, axis=0)
print(classid, " ", end="")
return x_augment, y_augment
In [12]:
%%time
if save_augmented_data:
D = Counter(y_train)
max_count = max(D.values()) * 2
for classid, count in D.items():
target = max_count-count
if target > 0:
x_augment, y_augment = augment_class(X_train, y_train, classid, target)
with open("augmented_data/" + str(classid) + "x", mode='wb') as f:
np.save(f, x_augment)
with open("augmented_data/" + str(classid) + "y", mode='wb') as f:
np.save(f, np.array(y_augment))
Augmented Data Preview¶
In [13]:
%%time
import os.path
fig, axes = plt.subplots(rows, columns, figsize=figsize,
subplot_kw={'xticks': [], 'yticks': []})
fig.subplots_adjust(hspace=0.3, wspace=0.10)
for classid in range(n_classes):
file_path = "augmented_data/" + str(classid) + "x"
if os.path.exists(file_path):
with open( file_path, mode='rb') as f:
x_augment = np.load(f)
with open("augmented_data/" + str(classid) + "y", mode='rb') as f:
y_augment = np.load(f)
X_train = np.append(X_train, x_augment, axis=0)
y_train = np.append(y_train, y_augment, axis=0)
axes.flat[classid].imshow(x_augment[0], interpolation='lanczos', cmap='viridis')
axes.flat[classid].set_title(sign_names[classid], fontsize=fontsize)
plt.show()
In [14]:
X_train, y_train = shuffle(X_train, y_train)
New Training Data Summary after Augmentation¶
In [15]:
n_train = len(X_train)
image_shape = X_train[0].shape
n_classes = len(set(y_train))
print("Number of training examples =", n_train)
print("Image data shape =", image_shape)
print("Number of classes =", n_classes)
In [16]:
fig, ax = plt.subplots(figsize=(20,5))
D = Counter(y_train)
plt.xticks(range(len(D)), D.keys())
plot = plt.bar(range(len(D)), D.values(), color="red", alpha=0.8, label='New Training')
In [17]:
import tensorflow as tf
EPOCHS = 90
BATCH_SIZE = 1024
Model Architecture¶
In [18]:
def variable_summaries(var):
"""Attach a lot of summaries to a Tensor (for TensorBoard visualization)."""
with tf.name_scope('summaries'):
mean = tf.reduce_mean(var)
tf.summary.scalar('mean', mean)
with tf.name_scope('stddev'):
stddev = tf.sqrt(tf.reduce_mean(tf.square(var - mean)))
tf.summary.scalar('stddev', stddev)
tf.summary.scalar('max', tf.reduce_max(var))
tf.summary.scalar('min', tf.reduce_min(var))
tf.summary.histogram('histogram', var)
In [19]:
from tensorflow.contrib.layers import flatten
def LeNet_Fast(x):
# Arguments used for tf.truncated_normal, randomly defines variables for the weights and biases for each layer
mu = 0
sigma = 0.1
with tf.name_scope('dropout_in'):
x_drop = tf.nn.dropout(x, keep_prob_in)
with tf.name_scope("Convolutional_1"):
with tf.name_scope('weights'):
# SOLUTION: Layer 1: Convolutional. Input = 32x32x3. Output = 28x28x6.
conv1_W = tf.Variable(tf.truncated_normal(shape=(5, 5, use_n_channels, 6), mean = mu, stddev = sigma),
name="conv1_weight")
variable_summaries(conv1_W)
with tf.name_scope('biases'):
conv1_b = tf.Variable(tf.zeros(6), name="conv1_bias")
variable_summaries(conv1_b)
conv1 = tf.nn.conv2d(x_drop, conv1_W, strides=[1, 1, 1, 1], padding='VALID') + conv1_b
# SOLUTION: Activation.
conv1 = tf.nn.relu(conv1)
# SOLUTION: Pooling. Input = 28x28x6. Output = 14x14x6.
conv1 = tf.nn.max_pool(conv1, ksize=[1, 2, 2, 1], strides=[1, 2, 2, 1], padding='VALID', name='conv1')
with tf.name_scope('dropout_conv'):
conv1_drop = tf.nn.dropout(conv1, keep_prob_conv)
with tf.name_scope("Convolutional_2"):
with tf.name_scope('weights'):
# SOLUTION: Layer 2: Convolutional. Output = 10x10x16.
conv2_W = tf.Variable(tf.truncated_normal(shape=(5, 5, 6, 16), mean = mu, stddev = sigma),
name="conv2_weight")
variable_summaries(conv2_W)
with tf.name_scope('biases'):
conv2_b = tf.Variable(tf.zeros(16), name="conv2_bias")
variable_summaries(conv2_b)
conv2 = tf.nn.conv2d(conv1_drop, conv2_W, strides=[1, 1, 1, 1], padding='VALID') + conv2_b
# SOLUTION: Activation.
conv2 = tf.nn.relu(conv2)
# SOLUTION: Pooling. Input = 10x10x16. Output = 5x5x16.
conv2 = tf.nn.max_pool(conv2, ksize=[1, 2, 2, 1], strides=[1, 2, 2, 1], padding='VALID', name='conv2')
with tf.name_scope("flatten"):
# SOLUTION: Flatten. Input = 5x5x16. Output = 400.
fc0 = flatten(conv2)
with tf.name_scope("FullyConnected_1"):
# SOLUTION: Layer 3: Fully Connected. Input = 400. Output = 120.
fc1_W = tf.Variable(tf.truncated_normal(shape=(400, 120), mean = mu, stddev = sigma))
fc1_b = tf.Variable(tf.zeros(120))
fc1 = tf.matmul(fc0, fc1_W) + fc1_b
# SOLUTION: Activation.
fc1 = tf.nn.relu(fc1)
with tf.name_scope("FullyConnected_2"):
# SOLUTION: Layer 4: Fully Connected. Input = 120. Output = 84.
fc2_W = tf.Variable(tf.truncated_normal(shape=(120, 84), mean = mu, stddev = sigma))
fc2_b = tf.Variable(tf.zeros(84))
fc2 = tf.matmul(fc1, fc2_W) + fc2_b
# SOLUTION: Activation.
fc2 = tf.nn.relu(fc2)
with tf.name_scope('dropout_fc'):
fc2_drop = tf.nn.dropout(fc2, keep_prob_fc)
with tf.name_scope("FullyConnected_3"):
# SOLUTION: Layer 5: Fully Connected. Input = 84. Output = 43.
fc3_W = tf.Variable(tf.truncated_normal(shape=(84, 43), mean = mu, stddev = sigma))
fc3_b = tf.Variable(tf.zeros(43))
logits = tf.matmul(fc2_drop, fc3_W) + fc3_b
return logits
In [20]:
from tensorflow.contrib.layers import flatten
def LeNet_Deep(x):
# Arguments used for tf.truncated_normal, randomly defines variables for the weights and biases for each layer
mu = 0
sigma = 0.1
with tf.name_scope('dropout_in'):
x_drop = tf.nn.dropout(x, keep_prob_in)
with tf.name_scope("Convolutional_1"):
with tf.name_scope('weights'):
# SOLUTION: Layer 1: Convolutional. Input = 32x32x3. Output = 28x28x32.
conv1_W = tf.Variable(tf.truncated_normal(shape=(5, 5, use_n_channels, 32),
mean = mu, stddev = sigma), name="conv1_weight")
variable_summaries(conv1_W)
with tf.name_scope('biases'):
conv1_b = tf.Variable(tf.zeros(32), name="conv1_bias")
variable_summaries(conv1_b)
conv1 = tf.nn.conv2d(x_drop, conv1_W, strides=[1, 1, 1, 1], padding='VALID') + conv1_b
# SOLUTION: Activation.
conv1 = tf.nn.relu(conv1)
# SOLUTION: Pooling. Input = 28x28x32. Output = 14x14x32.
conv1 = tf.nn.max_pool(conv1, ksize=[1, 2, 2, 1], strides=[1, 2, 2, 1], padding='VALID', name='conv1')
with tf.name_scope('dropout_conv'):
conv1_drop = tf.nn.dropout(conv1, keep_prob_conv)
with tf.name_scope("Convolutional_2"):
with tf.name_scope('weights'):
# SOLUTION: Layer 2: Convolutional. Output = 10x10x64.
conv2_W = tf.Variable(tf.truncated_normal(shape=(5, 5, 32, 64), mean = mu, stddev = sigma),
name="conv2_weight")
variable_summaries(conv2_W)
with tf.name_scope('biases'):
conv2_b = tf.Variable(tf.zeros(64), name="conv2_bias")
variable_summaries(conv2_b)
conv2 = tf.nn.conv2d(conv1_drop, conv2_W, strides=[1, 1, 1, 1], padding='VALID') + conv2_b
# SOLUTION: Activation.
conv2 = tf.nn.relu(conv2)
# SOLUTION: Pooling. Input = 10x10x64. Output = 5x5x64.
conv2 = tf.nn.max_pool(conv2, ksize=[1, 2, 2, 1], strides=[1, 2, 2, 1], padding='VALID', name='conv2')
with tf.name_scope("flatten"):
# SOLUTION: Flatten. Input = 5x5x64. Output = 1600.
fc0 = flatten(conv2)
with tf.name_scope("FullyConnected_1"):
# SOLUTION: Layer 3: Fully Connected. Input = 1600. Output = 120.
fc1_W = tf.Variable(tf.truncated_normal(shape=(1600, 120), mean = mu, stddev = sigma))
fc1_b = tf.Variable(tf.zeros(120))
fc1 = tf.matmul(fc0, fc1_W) + fc1_b
# SOLUTION: Activation.
fc1 = tf.nn.relu(fc1)
with tf.name_scope("FullyConnected_2"):
# SOLUTION: Layer 4: Fully Connected. Input = 120. Output = 84.
fc2_W = tf.Variable(tf.truncated_normal(shape=(120, 84), mean = mu, stddev = sigma))
fc2_b = tf.Variable(tf.zeros(84))
fc2 = tf.matmul(fc1, fc2_W) + fc2_b
# SOLUTION: Activation.
fc2 = tf.nn.relu(fc2)
with tf.name_scope('dropout_fc'):
fc2_drop = tf.nn.dropout(fc2, keep_prob_fc)
with tf.name_scope("FullyConnected_3"):
# SOLUTION: Layer 5: Fully Connected. Input = 84. Output = 43.
fc3_W = tf.Variable(tf.truncated_normal(shape=(84, 43), mean = mu, stddev = sigma))
fc3_b = tf.Variable(tf.zeros(43))
logits = tf.matmul(fc2_drop, fc3_W) + fc3_b
return logits
In [21]:
with tf.name_scope('input'):
x = tf.placeholder(tf.float32, (None, 32, 32, use_n_channels), name="x")
y = tf.placeholder(tf.int32, (None), name="labels")
with tf.name_scope('dropout_in'):
keep_prob_in = tf.placeholder(tf.float32, name="keep_prob_in")
with tf.name_scope('dropout_conv'):
keep_prob_conv = tf.placeholder(tf.float32, name="keep_prob_conv")
with tf.name_scope('dropout_fc'):
keep_prob_fc = tf.placeholder(tf.float32, name="keep_prob_fc")
one_hot_y = tf.one_hot(y, 43)
Train, Validate and Test the Model¶
A validation set can be used to assess how well the model is performing. A low accuracy on the training and validation sets imply underfitting. A high accuracy on the training set but low accuracy on the validation set implies overfitting.
In [22]:
rate = 0.001
logits = LeNet_Deep(x)
with tf.name_scope('cross_entropy'):
cross_entropy = tf.nn.softmax_cross_entropy_with_logits(labels=one_hot_y, logits=logits)
with tf.name_scope("total"):
loss_operation = tf.reduce_mean(cross_entropy)
tf.summary.scalar('mean', loss_operation)
with tf.name_scope("train"):
optimizer = tf.train.AdamOptimizer(learning_rate = rate)
training_operation = optimizer.minimize(loss_operation)
In [23]:
with tf.name_scope("accuracy"):
correct_prediction = tf.equal(tf.argmax(logits, 1), tf.argmax(one_hot_y, 1))
accuracy_operation = tf.reduce_mean(tf.cast(correct_prediction, tf.float32))
accuracy = tf.reduce_mean(tf.cast(correct_prediction, tf.float32))
tf.summary.scalar('accuracy', accuracy_operation)
saver = tf.train.Saver()
# returns accuracy
def evaluate(X_data, y_data, index=-1):
num_examples = len(X_data)
total_accuracy = 0
sess = tf.get_default_session()
for offset in range(0, num_examples, BATCH_SIZE):
batch_x, batch_y = X_data[offset:offset+BATCH_SIZE], y_data[offset:offset+BATCH_SIZE]
summary, accuracy = sess.run([merged, accuracy_operation], feed_dict={x: batch_x, y: batch_y,
keep_prob_fc: 1., keep_prob_in: 1.,
keep_prob_conv: 1.})
total_accuracy += (accuracy * len(batch_x))
if index >= 0:
train_writer.add_summary(summary, index)
else:
test_writer.add_summary(summary, index)
return total_accuracy / num_examples
#returns flatten array of predictions
def get_predictions(X_data, y_data):
num_examples = len(X_data)
total_accuracy = 0
sess = tf.get_default_session()
predictions = []
for offset in range(0, num_examples, BATCH_SIZE):
batch_x, batch_y = X_data[offset:offset+BATCH_SIZE], y_data[offset:offset+BATCH_SIZE]
softmax = sess.run(tf.nn.softmax(logits), feed_dict={x: batch_x, y: batch_y, keep_prob_fc: 1.,
keep_prob_in: 1., keep_prob_conv: 1.})
test_preds = [np.argmax(row) for row in softmax]
predictions.append(test_preds)
return sum(predictions, [])
In [24]:
with tf.Session() as sess:
merged = tf.summary.merge_all()
train_writer = tf.summary.FileWriter("log" + '/train', sess.graph)
test_writer = tf.summary.FileWriter("log" + '/test')
In [25]:
%%time
with tf.name_scope('train'):
train_step = tf.train.AdamOptimizer(learning_rate = rate).minimize(
cross_entropy)
with tf.Session() as sess:
sess.run(tf.global_variables_initializer())
num_examples = len(X_train)
print("Training...")
print()
for i in range(EPOCHS):
X_train, y_train = shuffle(X_train, y_train)
for offset in range(0, num_examples, BATCH_SIZE):
end = offset + BATCH_SIZE
batch_x, batch_y = X_train[offset:end], y_train[offset:end]
summary, _ = sess.run([merged, train_step], feed_dict={x: batch_x, y: batch_y, keep_prob_fc: 0.5,
keep_prob_in: 0.99, keep_prob_conv: 0.9})
train_writer.add_summary(summary, i)
validation_accuracy = evaluate(X_valid, y_valid, i)
print("EPOCH {} ...".format(i+1))
print("Validation Accuracy = {:.3f}".format(validation_accuracy))
print()
saver.save(sess, './lenet')
traning_accuracy = evaluate(X_train, y_train)
print("Training Accuracy = {:.3f}".format(traning_accuracy))
In [26]:
with tf.Session() as sess:
saver.restore(sess, tf.train.latest_checkpoint('.'))
with tf.name_scope('test'):
traning_accuracy = evaluate(X_train, y_train)
print("Training Accuracy = {:.3f}".format(traning_accuracy))
Test Accuracy¶
In [27]:
with tf.Session() as sess:
saver.restore(sess, tf.train.latest_checkpoint('.'))
test_accuracy = evaluate(X_test, y_test)
print("Test Accuracy = {:.3f}".format(test_accuracy))
predictions = get_predictions(X_test, y_test)
In [28]:
train_writer.close()
test_writer.close()
Detailed Analysis of Model Performance¶
Precision¶
In [29]:
from sklearn.metrics import precision_recall_fscore_support
precision,recall, fbeta_score, _ = precision_recall_fscore_support(y_test, predictions, average=None)
fig, ax = plt.subplots(figsize=(20,5))
plt.xticks(range(n_classes), range(n_classes))
plot = plt.bar(range(n_classes), precision, color="red", alpha=0.8, label='precision')
Recall¶
In [30]:
fig, ax = plt.subplots(figsize=(20,5))
plt.xticks(range(n_classes), range(n_classes))
plot = plt.bar(range(n_classes), recall, color="green", alpha=0.8, label='Recall')
Fβ score¶
In [31]:
fig, ax = plt.subplots(figsize=(20,5))
plt.xticks(range(n_classes), range(n_classes))
plot = plt.bar(range(n_classes), fbeta_score, color="blue", alpha=0.8, label='Recall')
Confusion Matrix¶
In [32]:
# Adopted from: http://scikit-learn.org/stable/_downloads/plot_confusion_matrix.py
import itertools
import numpy as np
import matplotlib.pyplot as plt
def plot_confusion_matrix(cm, classes,
normalize=False,
title='Confusion matrix',
cmap=plt.cm.Blues):
"""
This function prints and plots the confusion matrix.
Normalization can be applied by setting `normalize=True`.
"""
if normalize:
cm = cm.astype('float') / cm.sum(axis=1)[:, np.newaxis]
print("Normalized confusion matrix")
else:
print('Confusion matrix, without normalization')
plt.imshow(cm, interpolation='nearest', cmap=cmap)
plt.title(title)
plt.colorbar(orientation='horizontal')
tick_marks = np.arange(len(classes))
plt.xticks(tick_marks, classes, rotation=90)
plt.yticks(tick_marks, classes)
fmt = '.2f' if normalize else 'd'
thresh = cm.max() / 2.
for i, j in itertools.product(range(cm.shape[0]), range(cm.shape[1])):
plt.text(j, i, format(cm[i, j], fmt),
horizontalalignment="center",
color="white" if cm[i, j] > thresh else "black")
plt.tight_layout()
plt.ylabel('True label')
plt.xlabel('Predicted label')
In [33]:
from sklearn.metrics import confusion_matrix
c_matrix = confusion_matrix(y_test, predictions)
plt.figure(figsize=(25,25))
plot_confusion_matrix(c_matrix, classes=sign_names.values())
In [34]:
plt.figure(figsize=(25,25))
plot_confusion_matrix(c_matrix, classes=sign_names.values(), normalize=True)
Step 3: Test a Model on New Images¶
To give yourself more insight into how your model is working, download at least five pictures of German traffic signs from the web and use your model to predict the traffic sign type.
Load and Output the Images¶
In [35]:
from PIL import Image
from scipy import misc
import numpy
import glob
imageFolderPath = 'my_own_images/'
imagePath = glob.glob(imageFolderPath + '/*.png')
X_myown = numpy.array( [numpy.array(misc.imread(img, mode='RGB')) for img in imagePath] )
Y_myown = numpy.array( [int(path.split("/")[1].split(".")[0]) for path in imagePath])
In [36]:
count = len(X_myown)
columns = count
rows = 1
fig, axes = plt.subplots(rows, columns, figsize=figsize,
subplot_kw={'xticks': [], 'yticks': []})
fig.subplots_adjust(hspace=0.3, wspace=0.10)
indices = range(count)
for ax, i in zip(axes.flat, indices):
ax.imshow(X_myown[i], interpolation='lanczos')
ax.set_title(sign_names[Y_myown[i]], fontsize=fontsize)
plt.show()
Pre-Processing¶
In [37]:
%%time
X_myown = preprocess_gpu(X_myown)
In [38]:
fig, axes = plt.subplots(rows, columns, figsize=figsize,
subplot_kw={'xticks': [], 'yticks': []})
fig.subplots_adjust(hspace=0.3, wspace=0.10)
indices = range(count)
for ax, i in zip(axes.flat, indices):
ax.imshow(X_myown[i], interpolation='lanczos')
ax.set_title(sign_names[Y_myown[i]], fontsize=fontsize)
plt.show()
Predict the Sign Type for Each Image¶
In [39]:
### Run the predictions here and use the model to output the prediction for each image.
### Make sure to pre-process the images with the same pre-processing pipeline used earlier.
### Feel free to use as many code cells as needed.
with tf.Session() as sess:
saver.restore(sess, tf.train.latest_checkpoint('.'))
test_accuracy = evaluate(X_myown, Y_myown)
print("Test Accuracy = {:.3f}".format(test_accuracy))
predictions = get_predictions(X_myown, Y_myown)
Analyze Performance¶
Output Top 5 Softmax Probabilities For Each Image Found on the Web¶
For each of the new images, print out the model's softmax probabilities to show the certainty of the model's predictions (limit the output to the top 5 probabilities for each image). tf.nn.top_k could prove helpful here.
In [41]:
with tf.Session() as sess:
saver.restore(sess, tf.train.latest_checkpoint('.'))
sf = sess.run(tf.nn.softmax(logits), feed_dict={x: X_myown, y: Y_myown, keep_prob_fc: 1.,
keep_prob_in: 1., keep_prob_conv: 1.})
topk = sess.run(tf.nn.top_k(tf.constant(sf), k=3))
In [42]:
fig, axes = plt.subplots(5, 4, figsize=figsize,
subplot_kw={'xticks': [], 'yticks': []})
fig.subplots_adjust(hspace=0.3, wspace=0.10)
x_label = ["Highest Probability", "2nd Highest", "3rd Highest"]
for index in indices:
axes[index][0].imshow(X_myown[index], interpolation='lanczos')
axes[index][0].set_ylabel(sign_names[Y_myown[index]], fontsize=11, color="grey")
axes[index][0].set_xlabel("Test Image", fontsize=11, color="blue")
for num, name in enumerate(topk.indices[index]):
ax = axes[index][num+1]
ax.imshow(X_train[numpy.where(y_train==name)[0][0]], interpolation='lanczos')
score = topk.values[index][num]
color = "green" if name == Y_myown[index] else "red" # green color for correct prediction
ax.set_title(str(score), fontsize=24, color=color)
ax.set_ylabel(sign_names[name], fontsize=11, color="grey")
ax.set_xlabel(x_label[num], fontsize=11)
plt.show()
Visualize the Neural Network's State with Test Images¶
This Section is not required to complete but acts as an additional excersise for understaning the output of a neural network's weights. While neural networks can be a great learning device they are often referred to as a black box. We can understand what the weights of a neural network look like better by plotting their feature maps. After successfully training your neural network you can see what it's feature maps look like by plotting the output of the network's weight layers in response to a test stimuli image. From these plotted feature maps, it's possible to see what characteristics of an image the network finds interesting. For a sign, maybe the inner network feature maps react with high activation to the sign's boundary outline or to the contrast in the sign's painted symbol.
Provided for you below is the function code that allows you to get the visualization output of any tensorflow weight layer you want. The inputs to the function should be a stimuli image, one used during training or a new one you provided, and then the tensorflow variable name that represents the layer's state during the training process, for instance if you wanted to see what the LeNet lab's feature maps looked like for it's second convolutional layer you could enter conv2 as the tf_activation variable.
For an example of what feature map outputs look like, check out NVIDIA's results in their paper End-to-End Deep Learning for Self-Driving Cars in the section Visualization of internal CNN State. NVIDIA was able to show that their network's inner weights had high activations to road boundary lines by comparing feature maps from an image with a clear path to one without. Try experimenting with a similar test to show that your trained network's weights are looking for interesting features, whether it's looking at differences in feature maps from images with or without a sign, or even what feature maps look like in a trained network vs a completely untrained one on the same sign image.
In [43]:
### Visualize your network's feature maps here.
### Feel free to use as many code cells as needed.
# image_input: the test image being fed into the network to produce the feature maps
# tf_activation: should be a tf variable name used during your training procedure that represents the calculated state of a specific weight layer
# activation_min/max: can be used to view the activation contrast in more detail, by default matplot sets min and max to the actual min and max values of the output
# plt_num: used to plot out multiple different weight feature map sets on the same block, just extend the plt number for each new feature map entry
def outputFeatureMap(image_input, tf_activation, activation_min=-1, activation_max=-1 ,plt_num=1):
# Here make sure to preprocess your image_input in a way your network expects
# with size, normalization, ect if needed
# image_input =
# Note: x should be the same name as your network's tensorflow data placeholder variable
# If you get an error tf_activation is not defined it may be having trouble accessing the variable from inside a function
activation = tf_activation.eval(session=sess,feed_dict={x : image_input, keep_prob_fc: 1.,
keep_prob_in: 1., keep_prob_conv: 1.})
featuremaps = activation.shape[3]
plt.figure(plt_num, figsize=(15,20))
for featuremap in range(featuremaps):
plt.subplot(8,8, featuremap+1) # sets the number of feature maps to show on each row and column
plt.title('FeatureMap ' + str(featuremap)) # displays the feature map number
if activation_min != -1 & activation_max != -1:
plt.imshow(activation[0,:,:, featuremap], interpolation="nearest", vmin =activation_min, vmax=activation_max, cmap="gray")
elif activation_max != -1:
plt.imshow(activation[0,:,:, featuremap], interpolation="nearest", vmax=activation_max, cmap="gray")
elif activation_min !=-1:
plt.imshow(activation[0,:,:, featuremap], interpolation="nearest", vmin=activation_min, cmap="gray")
else:
plt.imshow(activation[0,:,:, featuremap], interpolation="nearest", cmap="gray")
In [44]:
with tf.Session() as sess:
saver.restore(sess, tf.train.latest_checkpoint('.'))
for i in indices:
plt.figure(0, figsize=(2,2))
plt.imshow(X_myown[i], interpolation='lanczos')
test_accuracy = evaluate(X_myown[i:i+1], Y_myown[i:i+1])
tf_activation = sess.graph.get_tensor_by_name('Convolutional_1/conv1:0')
outputFeatureMap(X_myown[i:i+1], tf_activation, plt_num=1)
tf_activation2 = sess.graph.get_tensor_by_name('Convolutional_2/conv2:0')
outputFeatureMap(X_myown[i:i+1], tf_activation2, plt_num=2)
plt.show()
Detailed Look via TensorBoard¶
In [12]:
import matplotlib.pyplot as plt
import matplotlib.image as mpimg
import numpy as np
imageFolderPath = 'TensorBoard/'
imagePath = glob.glob(imageFolderPath + '/*.png')
for path in imagePath:
plt.figure(figsize=(20,20))
imgplot = plt.imshow(mpimg.imread(path))
plt.axis('off')
plt.show()
In [ ]: